rustytokens/src/discovery/mod.rs

//! Filesystem walk with smart filtering, producing the list of source files to index.

use std::path::{Path, PathBuf};

use anyhow::Result;

use crate::config::{SKIP_DIRS, language_for};

#[derive(Debug, Clone)]
pub struct SourceFile {
    /// Absolute path on disk.
    pub path: PathBuf,
    /// Path relative to the project root (forward slashes).
    pub rel: String,
    pub language: &'static str,
}

pub fn discover(project: &Path) -> Result<Vec<SourceFile>> {
    let project_abs = project
        .canonicalize()
        .unwrap_or_else(|_| project.to_path_buf());
    let mut builder = ignore::WalkBuilder::new(&project_abs);
    builder.standard_filters(true);
    builder.require_git(false);
    builder.filter_entry(|entry| {
        if entry.file_type().is_some_and(|t| t.is_dir()) {
            let name = entry.file_name().to_string_lossy();
            !SKIP_DIRS.iter().any(|d| *d == name)
        } else {
            true
        }
    });

    let mut out = Vec::new();
    for entry in builder.build() {
        let entry = match entry {
            Ok(e) => e,
            Err(_) => continue,
        };
        let path = entry.path();
        if entry.file_type().is_some_and(|t| t.is_dir()) {
            continue;
        }
        let Some(language) = language_for(path) else {
            continue;
        };
        let rel = path
            .strip_prefix(&project_abs)
            .map(|p| p.to_string_lossy().replace('\\', "/"))
            .unwrap_or_else(|_| path.to_string_lossy().replace('\\', "/"));
        out.push(SourceFile {
            path: path.to_path_buf(),
            rel,
            language,
        });
    }
    out.sort_by(|a, b| a.rel.cmp(&b.rel));
    Ok(out)
}